Challenging AI to a Math Duel was a Terrible Mistake

Поділитися
Вставка
  • Опубліковано 27 січ 2025

КОМЕНТАРІ •

  • @PapaFlammy69
    @PapaFlammy69  Рік тому +34

    *_What even is society... Let me know if you wanna see more of this series and if you've got specific video topic ideas, make sure to leave some comments down below too! :D_*
    Check out Brilliant to get a 30 day free trial of awesomeness! =D brilliant.org/FlammableMaths
    Check out my newest uploads over on @RockHardWoodDaddy ! =D
    ua-cam.com/video/uPLCnymscOE/v-deo.html
    My Merch! =D papaflammy.myteespring.co/
    Support my channel and become a Patron! =) www.patreon.com/mathable
    ChatGPT Playlist: ua-cam.com/video/QoiWOF76Bkc/v-deo.html

    • @oni8337
      @oni8337 Рік тому +1

      Glad to see you've made an awesome recovery :D

    • @table5584
      @table5584 Рік тому

      Flammable Maths

  • @AdondeBoy
    @AdondeBoy Рік тому +110

    2:50 flammy reading an exclamation mark as factorial proves he is extremely based

    • @instinx9154
      @instinx9154 Рік тому +3

      I knew I wasn't the only one who noticed that.

  • @mr.inhuman7932
    @mr.inhuman7932 Рік тому +176

    I am proud of this community. We all came together and got Papa Flammy the equipment he needs.

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +19

      Me too

    • @stevenm3914
      @stevenm3914 Рік тому +3

      yup it wasnt much but it was done through honest work.

    • @souravjha206
      @souravjha206 Рік тому +3

      I am late but why didn't he had 3-2-1 data back up?

  • @oscarbizard2411
    @oscarbizard2411 Рік тому +48

    2:51 hahaha only papa flammy could read "11!" as "eleven factorial"😂 love it

  • @Stdvwr
    @Stdvwr Рік тому +162

    e^x is approx. 1, fucking legendary engineering

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +19

      :D

    • @oni8337
      @oni8337 Рік тому +16

      applying the same zero order approximation to sin(x) you get that sin(x) = 0

    • @GoKotlinJava
      @GoKotlinJava Рік тому +4

      can you get better results by doing it like this by taking two cases?
      1. let x0, then solve by assuming e^x=x for example
      Am i right? I have no clue. I just saw the graph for e^x and started my own assumptions

    • @Smitology
      @Smitology Рік тому +4

      @@oni8337 x=0 is another useful zeroth order approximation

    • @oni8337
      @oni8337 Рік тому +1

      @@Smitology that's exactly what i just did

  • @matthewb2877
    @matthewb2877 Рік тому +142

    Throwing it out there: GPT 3.5 is EXTREMELY bad at math. If you want to continue this series (and honestly have it do anything than very basic addition), shelling out 20$ for GPT-4 would increase the value of these videos by a lot.
    Love your content!

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +47

      Will make sure to invest that small amount!

    • @HL-xi7sz
      @HL-xi7sz Рік тому +12

      Tried it with GPT-4 and it gives much more interesting problems!

    • @carlos999cod4
      @carlos999cod4 Рік тому +3

      Even that sucks at math

    • @jonathanfleming8500
      @jonathanfleming8500 Рік тому +9

      It couldn't solve simple trigonometry for me. Yet when I asked it a fairly complicated numeric differential equation; it got it first try.

    • @SorryBeeinLate
      @SorryBeeinLate Рік тому +9

      Also GPT 4 can now execute code, which adds all the computational power of python sci libs to the game, so if used to it's full potential it's actually not that bad. I use it as a study aid all the time.

  • @marciomilisse8074
    @marciomilisse8074 Рік тому +35

    let´s kick it up to eleven Factorial 💀

  • @soupisfornoobs4081
    @soupisfornoobs4081 Рік тому +23

    The last one with the roots is completely solvable, just set the expression under the cbrt to (a+1)^3 and the expression under the sqrt to a^2 and solve for x. It's just a cubic polynomial!

    • @soupisfornoobs4081
      @soupisfornoobs4081 Рік тому +5

      For those curious what that looks like here's a little bit of it
      x-1 = a² x=a²+1
      4x+9=(a+1)³ => 4a²+13 = (a+1)³
      Expanding our and simplifying we get
      a³-a²+3a-12=0, I haven't found how to solve it algebraically but the cubic formula gives the same answer Wolfram does, just solve for a then x = a²+1
      Keen readers will have noticed this method translates to all sums of 2 radicals, giving polynomials of degree equal to the highest root degree

    • @kepler-452b7
      @kepler-452b7 Рік тому +1

      @@soupisfornoobs4081Thank you ! You just thought me a new trick

    • @nanamacapagal8342
      @nanamacapagal8342 Рік тому

      ​@@soupisfornoobs4081 yeah I think cubic formula is the way to go here, here's how I did it:
      substitute u = a - 1/3 or a = u + 1/3
      you wind up with:
      u^3 + (8/3)u - 299/27 = 0.
      i will use this version of the cubic formula:
      if u^3 + 3pu + 2q = 0,
      u = cbrt(-q + sqrt(q^2 + p^3)) + cbrt(-q - sqrt(q^2 + p^3))
      some chugging through and simplifying later, this pops out:
      u = (1/3)(cbrt(299/2 + sqrt(91449)/2) + cbrt(299/2 - sqrt(91149)/2))
      reverse both substitutions u -> a -> x.
      x = ((cbrt(299/2 + sqrt(91449)/2) + cbrt(299/2 - sqrt(91149)/2))/3 + 1/3)^2 + 1
      which is roughly 5.70231. I ran newton's method on this and the results seem to agree.
      I did also run newton's method on the earlier [bullshit] transcendental question and I got the same result as wolfram.

    • @adandap
      @adandap Рік тому

      I was going to type the same thing. My approach was to write 4 x + 9 = 4 (x-1) + 13 then set u = sqrt(x-1). Then a little algebra gives u^3 - u^2 + 3u -12 =0 which can be solved exactly.

    • @encounteringjack5699
      @encounteringjack5699 Рік тому

      Nice. I put it into a cubic polynomial without using a’s. Ended up as,
      x^3 + 2(x^2) - 22x - 125 = 0

  • @Maths_3.1415
    @Maths_3.1415 Рік тому +2

    11:00 lol 😂

  • @ivolol
    @ivolol Рік тому +26

    I've heard that Khan Academy having been working with OpenAI to make a chat bot that is extremely more strict in answering math correctly, specifically so it can be used as a robot tutor. Might be cool to check that out if possible. Not sure if you'd have to ask them for access

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +6

      sounds gorgeous, let me check this one out!

    • @mikeschieffer2644
      @mikeschieffer2644 Рік тому +4

      @@PapaFlammy69 I believe they call theirs Khanmigo.

    • @SorryBeeinLate
      @SorryBeeinLate Рік тому +4

      GPT 4 with code execution is quite powerful as well.

  • @Physicsguy-r7k
    @Physicsguy-r7k Рік тому +2

    When mathematicians finally agree that they too do approximations all the time.🤯🤯

  • @mr.inhuman7932
    @mr.inhuman7932 Рік тому +26

    Seeing you solve problems "live" is actually fun. And ChatGPT makes for funny moments!!

  • @natebrown2805
    @natebrown2805 Рік тому +11

    i was inspired by this video to ask chatgpt for some integrals to solve, and when i asked it if i was right, in the process of solving, it claimed that 1+2/3-2/3=1-2/3=1/3

  • @jamespetersen212
    @jamespetersen212 Рік тому +3

    There is a way to get that last problem into a cubic which can then be solved to find an exact real solution. (4x+9)^(1/3)-sqrt(x-1)=1.
    Add sqrt(x-1) to both sides and cube both sides to get rid of the cube root. Expand the (1+sqrt(x-1))^3 on the right side to get
    4x+9=1+3sqrt(x-1)+3(x-1)+(x-1)sqrt(x-1). factor sqrt(x-1) and expand 3(x-1) to get 4x+9=3x-2+(x+2)sqrt(x-1). Bring 3x-2 to the left side to get
    x+11=(x+2)sqrt(x-1). Square to get (x+11)^2=(x+2)^2(x-1). Expand to get x^2+22x+121=(x^2+4x+4)(x-1). Multiply out the right side to get x^2+22x+121=x^3+3x^2-4. Subtract the left side from the equation to finally get 0=x^3+2x^2-22x-125. The real solution to this cubic solves (4x+9)^1/3-sqrt(x-1)=1.

  • @alexdefoc6919
    @alexdefoc6919 Рік тому +1

    Finally, now not only can students solve homework with chatgpt, but teachers can assign it for students and them solving with chatgpt and so on....we gonna be so dumb 🤣🤣

  • @Zacvh
    @Zacvh Рік тому +1

    11 factorial 😂 the sign of a true mathematician

  • @drgatsis
    @drgatsis Рік тому +3

    Hilarious! I noticed the train. I recently got the crank operated wheel coaster with the steel balls. So much fun to build with my nephew

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +2

      it is!!!

    • @drgatsis
      @drgatsis Рік тому +1

      @PapaFlammy69 so cool I found it randomly in a hobby shop here in Toronto

  • @knvids2812
    @knvids2812 Рік тому +3

    0:11 made me laugh for 2 minutes😂😂😂

  • @theOG2109
    @theOG2109 Рік тому +3

    Papa: Chatgtp could I have something algebraic solvable?
    Chatgtp: No! 😡

  • @kasel1979krettnach
    @kasel1979krettnach Рік тому +4

    bing AI was gaslighting and then kicking me out of the chat yesterday, just because it gave the wrong answer and I dared to ask it to calculate again...

  • @ternarycode
    @ternarycode Рік тому +1

    silly flammy, the ai knows once you bro, you never go

  • @theepicman8160
    @theepicman8160 Рік тому +3

    For the exponential equation which you had to solve, you could have solved it by keeping e^x on one side and the rest of the stuff on the other side and then analyze their domain and range. Another way was to graph both lhs and rhs and finding the intersection points by hit and trial.

  • @bartoszhipnarowicz7935
    @bartoszhipnarowicz7935 Рік тому +3

    Dude, because fo this video I understood chatgpt usefulness, love it

  • @juicetime910
    @juicetime910 Рік тому +1

    man is such a math guy when chat gpt put an exclamation point he said “11 factorial” 😂

  • @dcterr1
    @dcterr1 Рік тому

    I like how ChatGPT picked up on your badass lingo! That's rad to the max, dude!

  • @sierpinskibrot
    @sierpinskibrot Рік тому +2

    Any chance we could see some more ,you laugh you math' with Dotson ?? 🙏🙏 loved that series

  • @evelinastanchovska4171
    @evelinastanchovska4171 Рік тому +1

    On the second equation you have x*e^x, can you just use the lambert double u function and are there any nice way to express W(x+y)?

    • @Kondorgon
      @Kondorgon Рік тому

      You can, since x/(-3x^4+1)*e^x=1 is equivalent to x/(-3x^4+1)*e^(x/(-3x^4+1))=1 you get the equation 3x^4+x-W(1)=0, solving for x (using wolfram) you get the correct solutions

  • @timmyourking4482
    @timmyourking4482 Рік тому +5

    I've simplified the last question to be the cubic equation x^3+2x^2-22x-125=0, but it's still equally diabolic

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +2

      sucks dick, ngl

    • @nanamacapagal8342
      @nanamacapagal8342 Рік тому +1

      I was expecting degree 6 polynomials, at least cubics are slightly better
      I wound up with u^3 - u^2 + 3u - 12 = 0 with x = u^2 + 1. Slightly different but eh
      Then cubic formula.

  • @Wielorybkek
    @Wielorybkek Рік тому +5

    If you ask it more specific questions, like to give you some intégéral, then it's much nicer. would be great to see more videos like this in the area of human-computer interaction!

  • @DavidMcCoul
    @DavidMcCoul Рік тому +2

    This was totally rad and bro-tastic!

  • @DarthSidian
    @DarthSidian Рік тому +1

    I love how Papa Flammy also has a Razer mouse.

  • @Tower_Of_Chaos
    @Tower_Of_Chaos Рік тому +2

    Today's flammable maths non-trivial trivia: 3 is not equal to zero.

  • @Me-pt8sv
    @Me-pt8sv Рік тому +3

    The third problem could probably be solved by the lambert W function if you play with it enough

  • @marek9741
    @marek9741 Рік тому +1

    Veritasium made a great video on math duels! They used to be very popular among mathematics

  • @GearsScrewlose
    @GearsScrewlose Рік тому +4

    If you prove there aren't any integer solutions algebraically, you'll be halfway home to finding a real solution. What I mean is let 4x+9 = n^3, and x-1 = m^2, you can find a linear combination of those two such that x vanishes and use the fact that n-m = 1 to get a cubic polynomial you can show has no real solution. Solving that cubic polynomial and back subbing get you x, but it's not worth the effort.

    • @GearsScrewlose
      @GearsScrewlose Рік тому +1

      I meant no real int solution. If I recall n = 3.2685 or some shit

  • @jinks908
    @jinks908 Рік тому +1

    Pleeeassseeee tell me where you got those chalkboards

  • @shy_dodecahedron
    @shy_dodecahedron Рік тому +7

    AI just generates something that looks like adequate speech, and that's just enough for normal humans.

  • @CoolCatDoingAKickflip
    @CoolCatDoingAKickflip Рік тому +1

    Sigma man versus lambda AI. 😎

  • @ofekshochat9920
    @ofekshochat9920 Рік тому +1

    Great video and good luck!

  • @mrosskne
    @mrosskne Рік тому +2

    how many levels of irony is GPT on?

  • @ipcheng8022
    @ipcheng8022 Рік тому +1

    You asked, the machine delivered

  • @mooshiros7053
    @mooshiros7053 Рік тому +6

    I really hope his students know he runs the greatest math channel on youtube (except for 3b1b)

  • @hilaryleung07
    @hilaryleung07 Рік тому +2

    Now, now, now ask ChatGPT what is 1+1 and it would give out a wrong answer

  • @AdityaKumar-gv4dj
    @AdityaKumar-gv4dj Рік тому +6

    I think you should have told chat gpt to give a question that has rational solutions

  • @athu_vathu
    @athu_vathu Рік тому +1

    what is that desk toy pendulum thing called

  • @martinshoosterman
    @martinshoosterman Рік тому +1

    im curious what would happen jf you kept asking it for harder questions.

  • @pixel0818
    @pixel0818 Рік тому +2

    i'm not very experienced using newton's method, but maybe you could atleast approximate the solutions that way. not sure if that's possible though

  • @markojojic6223
    @markojojic6223 2 місяці тому

    I know Im late but for the last one can't you just move the square root bit to the other side, take to the power of 6 and yield a quadratic like equation after a lot of refactoring and squaring

  • @Wielorybkek
    @Wielorybkek Рік тому +2

    awesome that we managed to collect the entire sum!

  • @ericyeahbaby3875
    @ericyeahbaby3875 Рік тому +1

    "I don't believe that a lot of skaters are good at mathematics" is one of the stupidest thing I've heard from a smart person in a long time

  • @evionlast
    @evionlast Рік тому +1

    I'm a skater and I like math, mathematician have rejected me... 😂

  • @Maths_3.1415
    @Maths_3.1415 Рік тому +3

    Ask for challenge to chatgpt 4
    It gives ridiculously hard problems of integration

  • @mr.inhuman7932
    @mr.inhuman7932 Рік тому +4

    I wish you a lot of luck at the chiropractic.

  • @Avighna
    @Avighna Рік тому

    20:05 press on exact form, there is a nice solution :cringe:

  • @رضاشریعت
    @رضاشریعت Рік тому +1

    Hope no teacher figures this out for creating exams

  • @Larry640
    @Larry640 Рік тому +3

    I remember giving chatgpt a calc 2 exam once and it got like a low B

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +2

      nice.

    • @talkgb
      @talkgb Рік тому

      correction: you used chatgpt on your own calc 2 exam and got a low B.

  • @aweebthatlovesmath4220
    @aweebthatlovesmath4220 Рік тому +1

    6:31 i think you need to prove that

  • @HaniaTauqeer-c2k
    @HaniaTauqeer-c2k 6 місяців тому

    I’m sorry, but I cannot focus on your video- where did you get that twisty metal thing!!

  • @howardlam6181
    @howardlam6181 Рік тому +4

    Where did ChatGPT learn all the bro stuff haha

    • @axelnils
      @axelnils Рік тому +2

      Elon Musk hasn’t confirmed that Andrew Dotson’s comment section wasn’t part of the training data

    • @howardlam6181
      @howardlam6181 Рік тому

      @@axelnils Elon Musk quitted OpenAI long before ChatGPT 3 became a thing.

    • @That_One_Guy...
      @That_One_Guy... Рік тому

      @@axelnils What the fuk

  • @STA-3
    @STA-3 Рік тому +1

    Why didn't you use 'x= (-b ± sqrt(b² -4ac))/2 is it because it's easier to mess up?

    • @PapaFlammy69
      @PapaFlammy69  Рік тому

      just preference

    • @Dylinian
      @Dylinian Рік тому +1

      I remember being very confused while living in Germany why they kept using something called the p q formel. I’m very pleased to learn it is just the quadratic formula for incels. Thanks, broseidon!

  • @spideysg1163
    @spideysg1163 Рік тому +2

    Are you from Germany or United States? Because you have a German accent but your country listed in your about page is the U.S just curious btw this was a great video!

  • @xanderlastname3281
    @xanderlastname3281 Рік тому +5

    I'm disappointed that you didn't keep the bro mode at the max throughout the entire video
    [Edit]
    I mean the bro vibe was there but you toned it down

  • @pwd-tk9
    @pwd-tk9 Рік тому

    Mathematicians are not trained prompt engineers. No hard feelings.

  • @TheMrAineas1
    @TheMrAineas1 Рік тому +2

    Imagine not using the newton raphson method by hand 😒

  • @pneumaniac14
    @pneumaniac14 Рік тому

    you can write the equation in the form f(x)e^f(x) =g(x)e^g(x). and then solve f(x)=g(x). Depending on the functions you might yield more solutions than that, because xe^x isn't bijective, but we are engineers now, and we only care about solutions that have f(x)=g(x).
    Actually fuck me, the solutions have f(x)=/=g(x). I can't be bothered but if anyone knows a simple way of relating W_0(x) to W_-1(x), then you can solve the equation.

  • @SmileyIsAlone
    @SmileyIsAlone Рік тому +1

    Chatgpt did my hw it’s got most of them wrong so I switched to a different ai and no I don’t let chatgpt do all my hw and I do understand the math problem I only use it check my answers

  • @koled224
    @koled224 Рік тому +1

    Try out bing for one too
    Also the flammily is strong!

  • @ricardoparada5375
    @ricardoparada5375 Рік тому +2

    U should see that thing try and do analysis proofs lmao

  • @joyboricua3721
    @joyboricua3721 Рік тому +1

    Engineering is approximating everything

  • @Bobbel888
    @Bobbel888 Рік тому

    wolframalpha yields a real solution with cubic roots

  • @Gurogun
    @Gurogun Рік тому +1

    nice video

  • @mr.inhuman7932
    @mr.inhuman7932 Рік тому +2

    Bazinga.

  • @nohandle102
    @nohandle102 Рік тому

    Me just staring at the swinging sticks

  • @_DD_15
    @_DD_15 Рік тому

    Broseidon 😂😂

  • @Elliamy01
    @Elliamy01 Рік тому +2

    Hope you enjoy the xps :)

  • @MaJetiGizzle
    @MaJetiGizzle Рік тому +2

    Would recommend GPT-4 with the Wolfram Alpha plugin and/or the Code Interpreter for some really interesting math features.

  • @monkerud2108
    @monkerud2108 Рік тому +1

    you know some dude has gone 140 kph on a skateboard. brosidon

  • @yasyasmarangoz3577
    @yasyasmarangoz3577 Рік тому +1

    Eleven factorial 💀

  • @hemandy94
    @hemandy94 Рік тому

    When you read the 11 ! As 11 factorial lol

    • @hemandy94
      @hemandy94 Рік тому

      Physicists: Estimating is key.
      I love how you had to estimate due to lack of other alternate solution. I never would've considered such a simple skill

    • @PapaFlammy69
      @PapaFlammy69  Рік тому

      :)

  • @rhosymedra6628
    @rhosymedra6628 Рік тому +1

    lol this was hilarious

  • @chrisglosser7318
    @chrisglosser7318 Рік тому

    I literally guessed the answer to last on on the first try

  • @rajat23ranjan
    @rajat23ranjan Рік тому

    hello sir from India
    can you teach
    indefinite integral just like
    gn berman fiitjee module level
    application of derivatives 10+2 level
    definite integration
    probablity 10+2 level
    some manipulations in trigonometric equation 10+2 sl loney level

  • @aniruddhvasishta8334
    @aniruddhvasishta8334 Рік тому +2

    ChatGPT can't do math bro

  • @SmileyIsAlone
    @SmileyIsAlone Рік тому +1

    I’m surprised I find this entertaining

  • @turkinone2142
    @turkinone2142 Рік тому +3

    2:50 did he just read "11!" as a factorial 🤡

    • @PapaFlammy69
      @PapaFlammy69  Рік тому +2

      sure thing :^) Is there any other meaning to ! besides factorial?

  • @hajnasaju2390
    @hajnasaju2390 Рік тому

    Waiting for asmr video

  • @dcterr1
    @dcterr1 Рік тому

    You didn't solve the last problem in the best possible way. You need to first isolate the cube root on the left side and then cube both sides, which leaves an 4x + 9 on the left side and (3x - 2) + (x + 2) sqrt(3x - 1) on the right side. Now isolate the term with the square root on the right by subtracting 3x - 2 from both sides, leaving you with x + 11 on the left side and (x + 2) sqrt(3x - 1) on the right side. Now squaring both sides removes all radicals, leaving you with polynomials on each side. Finally, by bringing all polynomials terms to one side, we obtain the cubic equation x^3 + 2x^2 - 22x - 125 = 0. As is the case with all cubic equations, this equation yields exact solutions involving radicals, one of which is real in this case. So ChatGPT gave you a very difficult problem to solve, but not an impossible one! On the other hand, the previous equation cannot be solved by elementary methods and needs to be numerically approximated, unless there's a special function I don't know about that it can be expressed in terms of, but I doubt this is the case.

    • @Jimmy-vg2gd
      @Jimmy-vg2gd Рік тому +1

      I'm disappointed of him . This way to solve is obvious. Why he didn't use it😢

    • @dcterr1
      @dcterr1 Рік тому

      @@Jimmy-vg2gd Don't know. I guess it just didn't occur to him when he made the video. He should redo this one!

  • @chessyman6973
    @chessyman6973 9 місяців тому

    What bullshit

  • @merzoukimerzoukis2371
    @merzoukimerzoukis2371 Рік тому +1

    second comment

  • @kevinscheengsbier6130
    @kevinscheengsbier6130 Рік тому +1

    That series gonna be cool asf

  • @stevenscottoddballz
    @stevenscottoddballz Рік тому

    6:30 I never understood about A HAS to be 1. I got A=3, B=5 and C=-2, put the numbers and got -2 and 1/3. What is the reason for making A=1? ~
    20
    Σ {5R-20}
    R=0
    I have started learning about Sigma, but this one has me stumped.
    I have been using [n(a₁ + aₙ)] ÷ 2
    [20(-20+80)] ÷ 2 = 600
    But when I enter my answer of 600, I get a .
    Where is my error?

  • @itsprivate3061
    @itsprivate3061 Рік тому

    gigachad ChatGPT